This laboratory is improving quality management for experimental structures of biological molecules - using all-atom contact analysis (including the hydrogens) and updated geometrical and torsion-angle criteria to guide procedures for validating and correcting local problems in 3D structures of proteins and RNAs. Progress in the previous grant period yielded widespread acceptance of these new techniques, producing an observable improvement in quality measures across all new structures in the worldwide Protein Data Bank. There is now a need for further fundamental research to underpin the correctness of individual methodological details and extend its applicability to new classes of structures. A recent breakthrough was recognition of an error-producing systematic local ambiguity possible in one type of NMR data. These now can be analyzed to avoid faulty ensemble models. For assessment of template-based modeling in the CASP8 prediction experiment, measures were developed that go beyond the C1 backbone to assess the full predicted models, many of which are now accurate enough that such criteria are appropriate. A very important expansion would be to enable improvement of structural accuracy at the lower resolutions typical for the large molecular complexes whose structures are the most significant for biological and medical research. This will depend on analysis of the patterns of systematic distortion in low- resolution electron density, and modeling strategies that are not misled by those distortions. The relevance of this project for NIGMS is to increase the impact of a very large and important segment of the funded research, leading across-the-board to better biological understanding and better prospects for demanding applications such as drug design, arising from better knowledge of the 3D molecular conformations, their dynamics and chemistry, and their detailed interactions with ligand molecules. PUBLIC HEALTH RELEVANCE: The science behind our all-atom contact and MolProbity techniques has resulted in an effective 3D "spell-checker" for macromolecular crystal structures, using the underlying science and context sensitivity to recognize and correct systematic errors, but doing no harm analogous to an overzealous spell-checker changing "CASP" to "gasp". Over the past 4-year grant period, this system was adopted widely enough to produce an observable improvement in quality measures across new depositions to the worldwide Protein Data Bank;this improved accuracy is especially crucial for detail-sensitive biomedical research such as drug design. This grant proposes fundamental research to enhance that underlying science and extend its benefits to other "languages" of structure such as NMR methodology, homology modeling, the increasingly important RNAs, and the lower-resolution structures characteristic of the most biologically and medically important large molecular complexes.